Study roster power audit — 2026-09-09

The mobile roster buys less printed damage and Hull than gunline, but combines high Evasion, Speed above the ordinary Evasive movement threshold, and three free support actions. Weight alone therefore does not measure equivalent battlefield power. This audit identifies plausible mechanisms behind the preliminary mobile-versus-gunline result; it neither confirms that small-panel win rate nor recommends a live nerf.

Reproduction and scope

pnpm exec tsx cli/study-power-audit.ts docs/tuning/evidence/2026-09-09-balance-study/power-audit.json
pnpm typecheck

The command refuses to overwrite an existing output; choose a new output filename for a repeat. Retained evidence contains the complete cloned baseline catalog/rules, all six concrete study rosters, source SHA-256 fingerprints, per-mech and arm budgets, 5,940 ranged rows, 90 melee matchup rows, and conditional Lock On support values. No games or random samples were run, and no live data or frozen study code was changed. Typecheck passed; runtime checks validate builds, probability mass, independent hit enumeration against the existing damage helper, and the expected absence of a bare-shot change when free Lock On is deleted.

The matrix enumerates current base weapon dice against fresh stock Courier, Foreman, Bastion, Reaver and Lantern bodies, at range 4/8/12, clear/soft/hard cover, with attacker movement and target Evasive separately toggled. Out-of-range rows have no attack value. Current long-range difficulty is zero, so legal long range changes reach without changing the neutral hit calculation.

“Printed damage” integrates hit, random body location, hard-cover leg absorption, penetration/glances and base damage dice. It excludes critical-table extras, Heat consequences, damage caps, part transfer, early termination within a volley, loads, Overdrive, aimed locations, reactions, Enraged, CP and action selection. It is neither actual Hull removed nor expected kills. Initial melee probabilities exclude exchanges, casualties and subsequent unavailable opportunities. These limits especially understate energy/control effects and opportunity denial.

Hard budgets and body power

Each Courier uses 33/35 personal weight; each gunline Foreman 48/50; each dual-blade Reaver 46/50; each Lantern 42/50. Bastion uses 62/75. Thus mobile leaves 46 team weight unused but has only six total personal-cap weight available across its existing chassis. Unspent team weight is not directly convertible into more weapons on these three bodies; raising their weight without considering personal caps would change legality before equating power.

Courier Core has 9 Hull / Armor 4, Foreman 12 / 6, and Bastion 18 / 7. Bastion also has reactor capacity 11 versus Courier 5 and Foreman 6. Those are meaningful endurance tradeoffs, while low Evasion and Speed reduce avoidance and access. Hull totals should not be collapsed into generic effective HP: hit-location allocation, aimed Core attacks, armor, disable effects and movement all change their value.

Delivered fire versus printed fire

These are sums of each roster member’s best independent base Fire at range 4 against the same fresh Foreman target, with no support or movement conditions. They describe three reference shooting opportunities, not a legal simultaneous volley or a bot plan.

Gunline leads mobile in raw fire under common conditions. However, mobile moving against a non-Evasive Foreman retains 12.42 clear reference damage, while moving gunline against an Evasive stock Vex Courier produces 8.44. This is an intentionally asymmetric example involving different target defenses, not a matchup forecast. It illustrates why comparing weapon damage alone misses how mobility delivers attacks while changing incoming attacks. Ordinary Speed 6 can reach the Evasive threshold of five traveled hexes in one Move; Speed 4 cannot without another movement allowance or bonus. This threshold, access to objectives and movement penalties for heavy weapons are distinct axes worth isolating in the full-game study.

Control’s strong base damage also means it is not a pure support roster. Its Heat effects and jammer utility are additional dimensions omitted here. A family label is not proof of its actual strategy.

Pilot and free-back concentration

Pilot selection has no weight cost. Gunnery raises ranged accuracy; Piloting raises Evasion, melee accuracy, thresholded pool size and first-actor priority; Systems raises E-Def and Lock On success. Those effects are not interchangeable point for point. The audit retains baseline-versus-neutral pilot stats per mech and neutral-attacker shots against the unchanged baseline target.

For example, Sato’s Foreman Autocannon against a fresh Foreman in clear conditions hits 80% per projectile and produces 5.97 printed damage per Fire. Setting the attacker pilot stats to one gives 70% and 5.23. Ilse already has Gunnery one, so the same controlled attack does not change. Vex’s Courier Rifle instead improves from 70% to 75% when neutralized, while Vex loses one Evasion on defense. The existing neutralpilots full-game candidate changes both teams’ stats, abilities and bonds together; its result cannot identify a single stat’s marginal value.

Only the melee family activates special pilot abilities: Ruko’s Momentum and Marrow’s Riposte share a faction. All other study families have no qualifying faction/specialty partner. In particular, the three-Courier roster does not activate Vex’s Pace or Q’s Scramble; gunline does not activate Sato’s Deadeye. Bond partners are fielded in mixed/armored (Vex–Ilse), melee (Ruko–Marrow), and control (Q–Sato), but the bonus still requires the live partner within two hexes. The study does not yet represent several plausible ability-supported strategies.

All three mobile sensor masts supply free Lock On and Sensors 13. Vex Courier Rifle versus fresh Foreman rises from 70% to 86.67% hit chance conditional on an existing lock in clear conditions. Vex’s own Lock On succeeds 60% against that target, so one isolated free attempt before that attack raises expected hit chance to 80%, before accounting for later allied use. Under hard cover, the conditional change is smaller: 47.64% to 52.50%. Locks do not stack. Reach, line of sight, target choice and activation ordering determine actual value.

The nofreelock deletion leaves all unsupported shot values exactly unchanged, as the audit asserts. Its full-game effect therefore measures access/action economy and downstream behavior, not a hidden weapon stat. This is a particularly useful diagnostic of power not separately priced in the hard-weight system.

Melee opportunity economy

Against fresh Foreman with no conditions or rerolls:

These exact initial distributions use class, Piloting, handling and source-specific accuracy. “All-defense” spends the entire successful pool on defense; the same dice cannot also provide the displayed offensive output. Heavy strikes normally cost two opportunities to parry. An intact shield reduces that cost to one regardless of which successful opportunities are selected, explaining Bastion’s defensive efficiency. Melee’s large initial pools do not establish match strength: approach time, Full-action access, target selection, damage and reactive interruption remain decisive.

A practical power budget

Track a vector before inventing a single price: delivered offense (range/cover/Heat and action costs), survival (Evasion/armor/Core and disable risk), access (movement/terrain/Evasive threshold), exchange control (initial melee pool/priority/shields/reactions), and team utility (free actions, bonds, abilities, target locks). Retain hard equipment feasibility alongside this vector.

The strongest next controls are already in the study: free-lock deletion, neutral pilots, movement changes, ranged reliability and survival changes. Interpret each by its mechanism rather than selecting the highest aggregate win-rate flattening. Follow-up rosters should activate actual ability synergies and include matched pilot/body swaps; otherwise a weak “control” or “gunline” result could reflect the sampled composition rather than that strategy’s ceiling. Use full-game exposure telemetry to test whether extra movement becomes earlier damage, safer exchanges or more scoring. Do not rebalance mobile solely from its unused weight or the preliminary small-panel result.